Should ChatGPT bots be banned from the EEVblog Forum?

[EEVblog]

As the poll says.
Vote and discuss.

[xrunner]

I say yes - ban the BOTS. But do not ban responses generated by ChatGPT provided the member attributes them to ChatGPT.

[MT]

But what if ChatGPT knows the meaning of life!!? How could we humans possibly ban such deep knowledge!?

[EEVblog]

But do not ban responses generated by ChatGPT provided the member attributes them to ChatGPT.

Yes, that could become troublesome really quickly. If there was a rule that said no use of ChatGPT unless attributued then it's likely people will start reporting responses they think are ChatGPT.

[pcprogrammer]

I say yes - ban the BOTS. But do not ban responses generated by ChatGPT provided the member attributes them to ChatGPT.

I agree with that.

If someone wants to use ChatGPT to provide an answer that is fine, as long as they state the source and make it clear that it might be incorrect. Because these incorrect responses, like for instance the highlight functionality in Kicad, throws people of.

[xrunner]

But do not ban responses generated by ChatGPT provided the member attributes them to ChatGPT.

Yes, that could become troublesome really quickly. If there was a rule that said no use of ChatGPT unless attributued then it's likely people will start reporting responses they think are ChatGPT.

Yes I understand. It's a dilemma, it surely is. Let's see what the Greater Minds than mine say about it. 

[pcprogrammer]

But what if ChatGPT knows the meaning of life!!? How could we humans possibly ban such deep knowledge!?

Since these machines feed on available information and don't really think for them selves it is not likely that this will happen 

[vk6zgo]

How about dogs?
"On the internet nobody knows you're a dog!"

[GromBeestje]

I'd say ban them. They'll waste everyone's times.
Let's put it this way, if I wanted a bot to answer, I would ask a bot. I come to a forum because I want a human answer.

[MK14]

I'd say ban them. They'll waste everyone's times.
Let's put it this way, if I wanted a bot to answer, I would ask a bot. I come to a forum because I want a human answer.

That mirrors my feelings nicely, as well.

Current ChatGPT results, can have serious flaws and inaccuracies in it.  So, just posting unchecked / unmodified and unlabeled (as from ChatGPT) stuff, is very annoying and could waste peoples time, or even be dangerous.

[MT]

But what if ChatGPT knows the meaning of life!!? How could we humans possibly ban such deep knowledge!?

Since these machines feed on available information and don't really think for them selves it is not likely that this will happen 

Heeey! At your place humans about to ban farmers and farming, over at my place humans about to ban CO2 and go full bore CBDC's, who said humans really think! or think "better" then AI.

[wilfred]

How about dogs?
"On the internet nobody knows you're a dog!"

Did you hear about the agnostic dyslexic insomniac ?

He would lay awake at night wondering "Is there a dog?"

Edit: perhaps it's time he asked chatPTG

[MK14]

Speculation, on my part.

One worry, is that businesses, will use the ChatGPT / AI system (and similar, possibly via API's), to go round places, such as forums, and look for questions / threads / posts, related to their product range.  They will then get the bot's script, to automatically generate, an apparently useful and helpful, friendly reply, which looks like a typical genuine human post.
Except it will promote their product range.

Hypothetical example:

My hard disk has bad sector errors and seems to be faulty, help!

=> Your hard disk, does indeed seem to be faulty.  I'd suggest you get a new and much faster SSD.  The best, cheapest, most reliable, fastest, longest lasting and one you should get is:
Diskphoolery-SSD-infused-silver-mirror-flash-oxygen-free-delux SSD's (link provided ...).
They are a bit pricey at $25,000 but worth every penny.

[Nominal Animal]

I say yes - ban the BOTS. But do not ban responses generated by ChatGPT provided the member attributes them to ChatGPT.

Agreed.  I don't like ChatGPT (because it irks me to no end when anyone gives wrong advice), but I could definitely live with that.

As I see it, it is no different than when e.g. I myself mention up front that I'm a hobbyist in electronics, not an EE, so that the reader knows to consider the reliability of the answer.

[tom66]

If a law cannot be policed and enforced, it holds little value. Can you identify the usage of ChatGPT? It might be straightforward to detect if ChatGPT is generating all of a user's output, but would you be able to distinguish that this paragraph was composed by ChatGPT?

[MK14]

If a law cannot be policed and enforced, it holds little value. Can you identify the usage of ChatGPT? It might be straightforward to detect if ChatGPT is generating all of a user's output, but would you be able to distinguish that this paragraph was composed by ChatGPT?

I suspect, that if a user (for whatever reason).  Made many (genuine), human posts, here.  But sometimes (secretly) consulted with ChatGPT, and sometimes, either cut and pasted bits of its response(s) into their posts, or even rewrote them, to attempt to further hide, what is really going on.  We wouldn't be able, to easily tell.

So I agree, it would be difficult to reliably fully enforce, such rules.

It is a bit like trying to prevent someone using a calculator, when they are given online tests and exams.  It would be difficult to enforce in practice.

[pcprogrammer]

But what if ChatGPT knows the meaning of life!!? How could we humans possibly ban such deep knowledge!?

Since these machines feed on available information and don't really think for them selves it is not likely that this will happen 

Heeey! At your place humans about to ban farmers and farming, over at my place humans about to ban CO2 and go full bore CBDC's, who said humans really think! or think "better" then AI.

Yeah in my land of birth stupid things are happening.  But in France, where I live, they do stupid things too.

And definitely not my words that humans really think. Far from it, I feel proper thinking is far of for a large part of the human population. 

[Nominal Animal]

If a law cannot be policed and enforced, it holds little value. Can you identify the usage of ChatGPT? It might be straightforward to detect if ChatGPT is generating all of a user's output, but would you be able to distinguish that this paragraph was composed by ChatGPT?

I don't follow.  The idea is to ban accounts that post identifiable ChatGPT-generated content.  Users who post identifiable ChatGPT-generated content without mentioning will be told to stop, because it is annoying.  If you use ChatGPT in ways that make it unidentifiable, then you obviously filter its output (or modify it enough), and as a member take the responsibility for its content, just like you do for the content you actually created yourself.  No difference to e.g. copying the key points of your answer from somewhere else, is there?  It is kinda fake, and not okay, but not something we can police.

While usage of ChatGPT is not really identifiable, reliance on ChatGPT for content is identifiable, just like those who spout most common rules of thumb as the full complete truth.  Hopefully, other members will put such members to the task.  It does not matter how the content is generated to me, bad advice should always be pointed out and corrected.

As I see it, the ban on ChatGPT bots is a line set in the rules for interaction.  We all agree that theft is wrong, yet not all thieves will ever be caught.  Making theft illegal and punishable, we define the rules we try to operate by.  Similarly, by banning ChatGPT bots we are stating that we are not interested in bot-generated posts.  (Members such as myself will, however, ignore any member who starts relying too much on ChatGPT in their posts, even if they state clearly they do, simply because they contain little of interest to myself.  I find the same content doing a few quick online searches.)

So, while I do agree that selective application of a law is definitely unfair, I do not think it applies here, exactly; because all rules here are mediated by explicit human action by Dave or one of the moderators he trusts.  It is a softer, "behavioural rule", subject to discussion.

[Neutrion]

If a law cannot be policed and enforced, it holds little value. Can you identify the usage of ChatGPT? It might be straightforward to detect if ChatGPT is generating all of a user's output, but would you be able to distinguish that this paragraph was composed by ChatGPT?

I suspect, that if a user (for whatever reason).  Made many (genuine), human posts, here.  But sometimes (secretly) consulted with ChatGPT, and sometimes, either cut and pasted bits of its response(s) into their posts, or even rewrote them, to attempt to further hide, what is really going on.  We wouldn't be able, to easily tell.

So I agree, it would be difficult to reliably fully enforce, such rules.

It is a bit like trying to prevent someone using a calculator, when they are given online tests and exams.  It would be difficult to enforce in practice.

It will be not just this forum. If AI -human interactions on the internet won't be marked as created by an AI it will have such an effect on human communications as forged money to the financial system.
A total chaos, and degradation of human human interactions, and besically the end of the internet as we know it. You will not be able to trust anyone who you don't know personally.
Problem is, until this law will be created, a lot of valuable time will be lost. And that will be the case with every law which would try to regulate this whole issue.

But what if ChatGPT knows the meaning of life!!? How could we humans possibly ban such deep knowledge!?

Since these machines feed on available information and don't really think for them selves it is not likely that this will happen 

Well read this:
https://www.valaszonline.hu/2023/03/10/szegedy-krisztian-matematikus-google-mesterseges-intelligencia-interju/

Its in hungarian so please use some translator. It's an interview with one of the key AI developers at google, who is working on AI to understand complex mathematical problems and to solve them.
For people who try to calm themself just because the current things have some flaws.
Unfortunately he interviewer forgot to ask the main ethical philosophical questions, so we have to trust this guy and the likes that it will be everything OK because he thinks so...

[xrunner]

Check out this article on The Register -

Microsoft and GM deal means your next car might talk, lie, gaslight and manipulate you

'ChatGPT is going to be in everything' says automaker

https://www.theregister.com/2023/03/14/microsoft_gm_chatgpt_car/?td=rt-3a

[MK14]

It will be not just this forum. If AI -human interactions on the internet won't be marked as created by an AI it will have such an effect on human communications as forged money to the financial system.
A total chaos, and degradation of human human interactions, and besically the end of the internet as we know it. You will not be able to trust anyone who you don't know personally.
Problem is, until this law will be created, a lot of valuable time will be lost. And that will be the case with every law which would try to regulate this whole issue.

I agree with you.  But I think it will be just like a huge list of previous inventions, which caused dramatic changes in society and the way people live.

E.g. Internet search engines, such as google.  Previously, we had to seek out knowledgeable people,  find a suitable book or magazine, or find out by trial and error etc.

These days we can just google or similar, and rapidly find the answer to an amazing range of possible queries.

You may be right, and it could mess up the existing quality of (usually) human written material, out there.

But I think it is coming, and it would be very difficult to stop it, now.

EDIT:
Looking on the bright side, a similar / analogy, might be chess computers.  So, although chess computers (software), can beat even the best human chess players.  Many people, still enjoy playing chess.  Either against other humans or while playing chess computers / software.
The only real issue it has within chess, is that they have to be rather careful in chess competitions.  To make sure that none of the players are (cheating) by using chess software, to win.

[nctnico]

As the poll says.
Vote and discuss.

I voted for a total ban. There are enough websites with answers that are clearly AI generated and useless.

The only really usefull application of AI that I see would be to translate websites between any language while keeping the context. Like the auto-translator they have in Sci-fi series.

[tggzzz]

Has nobody noted that a forums get polluted by ChatGPT bots, the bots will ingest their own output as new training material. Does that remind anybody of positive feedback and echo chambers?

The only escape from that is if a chatbot can detect and discount a post because it is bot-speak.

[RoGeorge]

For now, there might still be ways to identify chat bots, for example by testing them for glitch tokens:



Identifiable or not, a big no from my side for any bots on EEVblog.  Bots must be banned from human networks.  GPT is no better than any email spam bot, not to mention the destructive potential if ChatGPT is weaponized as a propaganda bot.  Propaganda is no joke, can inflict more damage than a nuke.

Because of this danger (propaganda bots), I bet it won't be long until we will see laws regarding total ban of chat bots from social networks, or at least some sort of human-user proof/signature/certificate.

[Neutrion]

It will be not just this forum. If AI -human interactions on the internet won't be marked as created by an AI it will have such an effect on human communications as forged money to the financial system.
A total chaos, and degradation of human human interactions, and besically the end of the internet as we know it. You will not be able to trust anyone who you don't know personally.
Problem is, until this law will be created, a lot of valuable time will be lost. And that will be the case with every law which would try to regulate this whole issue.

I agree with you.  But I think it will be just like a huge list of previous inventions, which caused dramatic changes in society and the way people live.

E.g. Internet search engines, such as google.  Previously, we had to seek out knowledgeable people,  find a suitable book or magazine, or find out by trial and error etc.

These days we can just google or similar, and rapidly find the answer to an amazing range of possible queries.

You may be right, and it could mess up the existing quality of (usually) human written material, out there.

But I think it is coming, and it would be very difficult to stop it, now.

EDIT:
Looking on the bright side, a similar / analogy, might be chess computers.  So, although chess computers (software), can beat even the best human chess players.  Many people, still enjoy playing chess.  Either against other humans or while playing chess computers / software.
The only real issue it has within chess, is that they have to be rather careful in chess competitions.  To make sure that none of the players are (cheating) by using chess software, to win.

I agree that it will be hard to stop it. Still we have to try if we think it is not good.

The logic, that we had big changes in the past, and we were able to accomodate to them hence this time it will be all right as well, is kind of a same as if you would have a house in a flood area, and just because it was not reached by the flood in the last 200 years you would conclude that it can not be reached.
This is not a logical argumentation as for how we could tacle the problem, this is just a vague hope.
(Don't forget, the scale of the first world war was the consequence of the industrial revolution, and possibly the rise of the communism as well.)
And I clearly do not even see the social-political-philosophical proposals to solve the issue now.
In the last 6-8 years, since this issue came up seriously i could not see any meaningful proposals for the solutions.

Key for the adaptation is whether we have time for it. Here I see that we don't. Could we adapt ourselfs in the last 20 years to mitigate climate change? I don't think so.